WG/99/015 Evaluation of pumice and scoria samples from East

advertisement
BRITISH GEOLOGICAL SURVEY
Mineralogy and Petrology
Technical Report
WGl99115
EVALUATION OF PUMICE AND SCORIA
SAMPLES FROM EAST AFRICA AS
LIGHTWEIGHT AGGREGATES
E J Evans, S J D Inglethorpe & P D Wetton
With contributions by D E Bailey, 2. Bongole, J. Kagasi & S J Mathers
Date
0 1112/99
Classification
Open
Geographical Index
East Africa, Kenya, Tanzania
Subject index
Pumice, scoria, lightweight aggregate
Bibliographic reference
E J Evans,S D J Inglethorpe, & P D Wetton, Mineralogy and Petrology Group
ONERC 1999 British Geological Survey, Keyworth, Nottingham. NGI 2 5GG. UK.
WGl99115
Version 1.0
0 1112/99
....................................................................................................................................
........................................................................................................
.................................................................................
...............................................................................................................................................
1. INTRODUCTION
1.1 Definition of pumice and scoria
1.2. Pumice and scoria as lightweight aggregates
1.3. Other industrial uses of pumice and scoria ....................................................................................
2. METHODS
2.1. Chemistry ..........................................................................................................................................
2.2.1. Major element x-ray fluorescence (XRF) analysis .....................
2.2.2. Electron probe tnict-o-analysis (EPMA) .....................................
2.2. Mineralogy ........................................................................................
2.3. Petrography ...........................................................................................................................................
2.3. I . Thin section nzicroscopy
) ..............................
....................................................................
2.4. Physical properties
............................................
3
3
3
4
5
s
6
.............
7
2.4.4. Aggregate itnpact value (AIV)....................
...................................................
...........................................
8
2.4.5. Water absorption and density determination.. ..................
2.4.6. Percentage voids content ..................
2.4.7. Specific gravity (SG)........................................
...........................................
3. RESULTS AND DISCUSSION
3.1. Chemistry
9
3.1.1. Major element x-raypuorescence (XRF) analysis .............
.................................
9
3.1.2. Electron probe micro-analysis (EPMA)
...................................
....................... 9
3.2. Mineralogy
10
3.2.1. Whole rock rnineralogy
........................................
...................................
3.3. Petrography
...............................
3.3.1. S M K 3 - airfnll pumice ........................
....................... 10
3.3.2. S M K S - pumice
................. I I
3.3.3. SMK6 - muiella pumice
...............................................................................................................
..........................................................................................................................................
......................................................................................................................................
.............................................
...........................
...........................
..........................
12
.....................
3.4.2. unit weight,
3.4.3. Specific gravity .........................
3.4.4. Aggregate impact value (AIV) ..........
.....................................................................................................................................
.......................................................................................................................................
...................................................................................................................
.........................................................
..........................................................................................
...................................
...........................................
...........................................................
..............................................................................
..................................................................................................
.............................................................................................
4. CONCLUSIONS
15
5. RECOMMENDATIONS .......................................................................................................................
16
6. REFERENCES
16
7. ACKNOWLEDGEMENTS
18
TABLES
Table 1. Sample list and descriptions as supplied by the collector
19
Table 2. Uses of pumice (after Geitgey, 1994)
20
Table 3. Whole rock chemistry for the samples and some commercial pumices
21
Table 4. Classification of the samples based on SiOz % (after Best, 1982)
22
Table 5. Electron probe micro analysis of the glass composition
23
Table 6. Summary of XRD results ............................................................................................................
24
Table 7. Descriptive data from optical petrography
.24
Table 8. Grading frequency distribution
25
Table 9. Particle-size cumulative frequency
26
Table 10. ASTM lightweight aggregate (maximum values for dry loose weight) .................................
27
Table 11. Variation of unit weight (kg/m3) vs. grading ..........................................................................
28
Table 12. ASTM test material requirements based on maximum particle size (mm) ...........................
29
Table 13. Summary of miscellaneous physical properties ....................................................................... 30
Table 14. ASTM test methods for lightweight aggregate ........................................................................
31
WG/99/15
Version 1.0
0 1/12/99
FIGURES
Figure 1 Si02-Na20+K20rock classification scheme (Cox et al., 1979)
32
Figure 2 AFM plot of the electron probe microanalysis data for the samples analysed
33
Figure 3 AFM plot of the average data for the samples obtained on the electron probe and for some
published igneous rock types
33
Figure 4 Annotated Xray diffraction (XRD) profile for pumice sample SMKS
34
Figure 5 Annotated X-ray diffraction (XRD) profile for scoria sample SMT3
34
Figure 6 Cumulative particle-size chart for the samples
35
Figure 7 Cumulative particle-size chart Envelopes for samples (black) are shown in comparison to
astm particle-size specifications for lightweight aggregate (red, blue and green)
36
.
.
.
.
.
.
.
...............................................
...........
......................................................................................................................
..................................
....................................
........................................................................
.
.................................
PLATES
Plate 1 Transmitted light photomicrograph of SMK3 showing typical glass structure and relatively
undeformed vesicles (field of view 3.2 mm)
37
Plate 2. Transmitted light photomicrograph of SMK3 showing poorly connected porosity (field of
view 1.6 mm)
37
Plate 3 Transmitted light photomicrograph of SMKS showing alkali feldspar (centre and right) and
glass/vesicle structure (field of view 3.2 mm)
38
Plate 4 BSEM photomicrograph showing octahedral fluorite crystal infiitrated by volcanic glass at
core (SMK5)
38
Plate 5 BSEM photomicrograph of sanidine grains within pumice matrix (SMKS)
39
Plate 6. BSEMphotomicrograph showing sanidine feldspar lath in alignment with pumice fabric
(SMK6)
39
Plate 7 BSEM photomicrograph showing large vesicles surounded by highly deformed fine vesicles
(SMK8)
40
Plate 8 Transmitted light photomicrograph of SMK8 showing altered feldspar phenocrysts (field of
view is 1.62 mm)
40
Plate 9 Transmitted light photomicrograph of SMK8 showing altered and fragmented feldspar (field
of view is 1.62 mm)
41
Plate 10. BSEMphotomicrograph of euhedral augite adjacent to tabular grain tentatively identified
as ferroaugite (SMT1)
41
Plate 11 Transmitted light photomicrograph of SMT2 showing typical structure with near opaque
glass and microphenocrysts (field of view is 10.5 mm)
42
Plate 12 BSEM photomicrograph showing very fine grained clinopyroxene crystallites within glassy
42
matrix Micro-phenocrysts present are augites (SMT2)
.
...............................................................................................
................................................................................................................................................
.
............................................................................................
.
.................................................................................................................................................
.
...........................
.........................................................................................................................................................
.
.........................................................................................................................................................
.
...........................................................................................................................................
.
......................................................................................................................................
.
.
.
.................................................................................................................................
.............................................................................
.........................................................................
n
WG/99/ IS
Version 1.0
0 I / 12/99
1. INTRODUCTION
This report describes the mineralogy, petrography, chemistry and physical properties of
pumice and scoria samples, as a means of evaluating their suitability as lightweight
aggregates. The samples were collected from Kenya by Steve Mathers (BGS) and John
Kagasi (Mines and Geological Department) and from Tanzania by Steve Mathers (BGS)
and Zacharia Bongole (Ministry of Energy and Minerals). The work has been carried out
as part of the Volcanics for Construction (VOLCON) research project funded by the UK’s
Department for International Development (DFID) under the Knowledge and Research
(KAR) programme.
1.1 Definition of pumice and scoria
Pumice is usually white or grey and is a highly vesiculated pyroclastic material which at
the moment of effusion was almost liquid. It has a froth-like appearance that results from
the sudden release of dissolved vapours on solidification.
Scoria is also a vesicular
pyroclastic material but is composed of heavier ferruginous lavas and is usually denser
and darker in colour than pumice (Lapidus, 1990).
1.2. Pumice and scoria as lightweight aggregates
Because of their lightweight nature and insulation properties, both pumice and scoria are
used as lightweight aggregates in concrete products, for structural concrete, plaster
aggregate and loose fill aggregate.
For example, in the USA 60% of the pumice
consumed is used as lightweight aggregate (Melendez, 1997). The lightweight nature of
pumice and scoria results in masonry blocks with specific weights of approximately
650 kg/m3 compared to 1800-2000 kg/m’ for normal bricks (Loughborough, 1991).
Pumice is lighter in weight and colour than scoria which is of greater strength and darker
coloration (Geitgey, 1994).
According to Robbins (1984), there are three important economic considerations
governing the use of lightweight aggregate. Firstly, it has been shown that structural
lightweight concrete is 25% lighter than ordinary concrete without any loss in
compressive strength. Secondly, thermal insulation properties are valuable and it has
been shown that heat losses can be reduced by up to 50% if lightweight concrete is
WG/99/15
0 1/12/99
Version 1.0
substituted. Thirdly, fire resistance is an important consideration as lightweight concrete
has some 20% more resistance to fire than normal concrete.
As a building aggregate, pumice is characterised by low unit weight, high transportation
costs, narrow markets and a susceptibility to substitution. However, pumice also has
several other advantages over non-lightweight materials.
Robbins (1984) noted that
pumice is:
easier to handle
requires a lower consumption of mortar
minimises the foundation requirements and supporting columns in multi-storey
buildings
decreases the reinforcement requirements in suspended floors
is resistant to fire, condensation, pests and mildew
reduces the bulk required to meet thermal and acoustic standards.
1.3. Other industrial uses of pumice and scoria
The end uses of pumice and scoria generally overlap although pumice has a wider range
of applications than scoria. In the United States the largest end use of pumice by volume
is as lightweight aggregate in concrete, this is followed by use in the abrasive industry
(Geitgey, 1994). However, cheaper alternatives to pumice include Fullers Earth and silica
flour. In fact, for nearly every end use of pumice and scoria there are generally a number
of natural or synthetic alternatives that can perform equally well. Therefore, the market is
highly competitive.
Other industrial uses of pumice include: use as absorbents, architectural applications, use
as fillers and use as filter media. Table 2 shows the uses, processing and essential
properties of pumice (after Geitgey, 1994). If it is too dense to be used as lightweight
aggregate, scoria can be substituted for sand and gravel aggregates. However, it must
meet all the required specifications for these applications. Scoria is also frequently used
in landscaping and is valued as an absorbent, particularly in gas cooking grills
(Geigty, 1994).
The suitability of pumice and scoria for a particular end-use is dependent on their physical
properties such as density, grain-size, grain-shape, toughness and friability. The chemical
Version 1.0
WG/99/ 15
0 1112/99
composition, particularly Si02 content is also important because a high Si02 content will
increase hardness and increase resistance to chemical attack (Robbins, 1984). The uses of
pumice are diverse but in all cases detailed chemical, mineralogical, petrographic and
physical properties must be assessed.
2. METHODS
2.1. Chemistry
2.2.1. Major element X-ray fluorescence ( X R F ) analysis
This was determined by X-ray fluorescence (XRF) using a Phillips PW 2400 X-ray
fluorescence spectrophotometer. Approximately 5 g of sample is dried for 24 hours at
105°C. Loss on ignition (LOI) is usually calculated from the weight loss of 1 g of sample
heated at 1050°C for one hour. However, due to problems of sintering and fusion the LOI
of the samples was instead calculated from heating for one hour at 850°C. Fused glass
beads were prepared by fusing 0.9 g of sample with 9 g of dried lithium tetraborate
(Li2B407) flux at approximately 1200°C in a muffle furnace. The melt obtained was
poured into a platinum casting dish. Lithium iodide was then added to all samples before
fusion to act as a releasing agent. Subsequently fourteen major elements were calculated
as oxides using a standard Philips calibration algorithm and theoretically generated alpha
coefficient corrections. The LOI values obtained represent the loss of volatiles from the
samples due to reactions such as carbonate decomposition, sulphide oxidation and the loss
of moisture and structural water.
2.2.2. Electron probe micro-analysis (EPMA)
Typical glass compositions were obtained using a Cambridge Instruments Microscan V
electron microprobe fitted with an Oxford Instruments AN10/55S energy dispersive X-ray
analyser (EDXA) system equipped with a beryllium window detector.
Analysis was
performed at 15 kV accelerating voltage with a specimen current of 4.5-5 nA for 60 s
live-time.
Elements analysed were Na, Mg, Al, Si, K, Ca, Ti, Cr, Mn, Fe with 0
calculated by stoichiometry. To ensure that the X-ray excitation volume during analysis
did not contain significant volumes of impregnated resin Cl was also included as this is a
WG/99/ I5
Version 1.0
01/12/99
major component of the resin used. Typical detection limits for the analytical conditions
used were approximately 0.1 wt%.
2.2. Mineralogy
A representative sub-sample of approximately 50 g was taken by cone and quartering
from the jaw-crushed material and ground in an automatic pestle and mortar for 10
minutes. This was then wet-micronised for 10 minutes in deionised water and dried at
55°C. The dried powder was disaggregated using an agate pestle and mortar before
backloading into a standard aluminium sample holder. XRD analysis was carried out
using a Philips PW 1700 series diffractometer equipped with a cobalt-target tube and
operating at 45kV and 40mA. Whole-rock samples were scanned from 3-50'20at
0.4S028/minute. Diffraction data were analysed using Philips X'Pert software coupled to
an International Centre for Diffraction Data (ICDD) database running on a Gateway
personal computer system.
2.3. PETROGRAPHY
2.3.1. Thin section microscopy
Blue dye, resin-impregnated, polished thin sections were made for each of the samples
(with the exception of SMT3 as this was very granular in nature). These were examined
by transmitted light microscopy. Standard petrographic descriptions were compiled.
2.3.2. Back-scattered electron microscopy (BSEM)
The polished thin sections were prepared for BSEM imaging (Goldstein, 1981) by coating
with approximately 20 nm of carbon using an Emitech K950L Turbo Evaporator. BSEM
investigations were performed using a LEO 435VP scanning electron microscope (SEM)
fitted with an Oxford Instruments ISIS 300 EDXA system equipped with thin window
X-ray detector allowing elements from B to U to be analysed. BSEM imaging was
performed under conditions of 20 kV accelerating voltage with 500 pA probe current.
Mineral identifications were based on qualitative EDXA observations.
WG/99/15
Version 1.0
01/12/99
2.3.3. Petrographic image analysis (PIA)
Total porosity for the samples was determined using a Kontron Elektronik KS400 image
analysis system. Images were obtained under automated stage scans using a LEO 435VP
SEM operating in BSEM imaging mode. Porosity estimates were determined for 16
images from each sample at a resolution of 4.02 ,um/pixel to give a total analysed area of
approximately 200 mm*.
2.4. Physical properties
2.4.1. Sieve analysis
This was determined in accordance with ASTM C136 using the following sieve apertures;
14, 10, 5 , 2.36, 1.18, 0.60, 0.30, 0.15 and 0.075 mm. Particle size was calculated both as
a frequency distribution (i.e. mass retained between each sieve aperture) and as a
cumulative distribution (i.e. mass percentage finer than each sieve aperture).
2.4.2. Unit weight (ungraded material)
Unit weight was carried out in accordance with ASTM C29. It is defined as the mass of
dry aggregate filling a specified container divided by the capacity of that container
expressed in kg/m’.
Unit weight is often called unit mass, or density, or bulk density,
however the term unit weight will be used throughout.
2.4.3. Unit weight (graded material)
The unit weight of individual size fractions was determined by placing a weighed portion
of each fraction in a measuring cylinder of suitable size and recording the volume
occupied. The unit weight was then calculated by dividing the volume by the weight of
the size fraction.
2.4.4. Aggregate impact value (AIV)
AIV was determined to measure the relative strength of the samples. AIV determinations
were carried out using apparatus conforming to the requirements of BS 812: Part 112:
1990.
WGl99115
Version 1.0
0 I/ 12/99
2.4.5. Wuter ubsorption und density determination
This was carried out in accordance with BS 812 : Part 2 : 1995 Methods of determination
of density Section 5.4. From this test, three measures of density were made and water
absorption was determined.
Firstly, particle density on an oven-dried basis was
determined. This is defined as ‘the ratio of the oven-dried mass of a sample of aggregate
to the volume it occupies in water including both internal sealed voids and water
accessible voids.’ Secondly, particle density on a saturated and surface dry basis was
determined. This is defined as ‘the ratio of the combined mass of a sample of aggregate
and the mass of the water in the water accessible voids to the volume it occupies in water
including both internal sealed voids and water accessible voids.’
Thirdly, apparent
particle density was determined. This is defined as ‘the ratio of the oven-dried mass of a
sample of aggregate to the volume it occupies in water including any internal sealed voids
but excluding any water-accessible voids.’ Lastly, water absorption was determined. This
is defined as ‘the increase in mass of a sample of aggregate due to the penetration of water
into the water-accessible voids of the oven-dried aggregate.’
2.4.6. Percentage voids content
The percentage voids content was calculated as outlined in BS 812 : Part 2 : 1995
Methods of determination of density Section 6.3. It is defined as a measure of ‘the air
filled spaces between the aggregate particles in the container.’
2.4.7. Specific gruvity (SG)
This was determined using a standard pycnometry method. A portion of ground sample
was placed in a previously weighed 25 cm3 pycnometer bottle, weighed again, then
thoroughly wetted with deionised water using an ultrasonic bath. After ensuring no air
bubbles were trapped, deionised water was added up to the 25 cm3 mark and the
pycnometer bottle weighed again. The pycnometer bottle was also weighed when filled
with deionised water only. The SG was then calculated using the following equation:
WG/99/ 15
Version 1.0
0 1/12/99
Where:
A = weight of the pycnometer bottle
B = weight of the pycnometer bottle and dry powder
C = weight of the pycnometer bottle, wetted powder and deionised water
D = weight of the pycnometer bottle and deionised water only
3. RESULTS AND DISCUSSION
3.1. Chemistry
3.1.1. Major element X-ray fluorescence ( X R F ) analysis
Major element compositions for the samples are given in Table 3. For comparison,
chemical data for commercial pumices are also listed. The major element compositions
of SMK3, SMKS, SMK6, SMK8 and SMTl are broadly comparable to those of
commercial pumice, although their total iron contents are markedly higher. Data for
commercial scoria are limited but they usually have Si02 contents between SO and 60 %
(Harben et al 1996). However, both SMT2 and SMT3 have lower Si02 values than these.
A petrographic classification of the samples based on their Si02 content is shown in
Table 4. In addition, the silica and total alkali contents (NaOz and K20) were used as
ordinates to plot the samples on a rock discrimination chart (Figure 1) using MinPet 2.02
software. According to this chart, the samples can be classified as follows; SMK3, SMKS
and SMK6 are rhyolites, SMK8 is a trachyte, SMTl is a phonolite, SMT2 is a phonolitic
nephelinite and SMT3 is a basanitehephrite.
3.1.2. Electron probe micro-analysis (EPMA
The average glass compositions for each sample are shown in Table S along with a
commercial sample for comparison. Unfortunately, SMT3 was too vesicular to analyse so
no results are given. For each sample, results of the individual analyses were plotted on a
triangular AFM diagram (Figure 2) to show the variation in glass composition. For
sample SMTl it is clearly evident that MgO content is variable. The other samples plot in
clusters indicating a homogenous glass composition.
For each sample, average
WG/99/ 15
Version 1.0
0 I/ 12/99
microprobe analyses were also plotted alongside some data for igneous rocks (Figure 3).
Generally, samples plot in the alkali corner of the triangle adjacent to rhyolites, phonolite
and dacite. However, the glass present in these East African samples appears to be
enriched in F e 0 and depleted in MgO relative to the igneous rocks plotted.
3.2. Mineralogy
3.2.I . Whole rock minemlogy
XRD results are summarised in Table 6. AI1 the samples are dominated by poorly ordered
volcanic glass (sideromelane) as indicated by the broad X-ray peak between 20 - 40 28.
O
Minor felsic minerals were identified in the pumice samples (see example for SMKS in
Figure 4). In contrast, the two scoria samples are typified by the presence of mafic phases
such as clinopyroxene and olivine (see example for SMT3 in Figure 6).
3.3. Petrography
Petrographic descriptions, total porosity and vesicle sizes are summarised in Table 7.
3.3.1. SMK3 - Airjiull pumice
Optical petrography showed this sample to contain relatively fresh glass with little
evidence of devitrification and no phenocrysts present (Plate 1). Porosity within the glass
is dominated by vesicles 0.1-0.5 mm in diameter with a number of larger vesicles up to
3 mm diameter also present. Vesicles show sub-spherical to ovoid forms, with limited
distortion of the original bubble shape in comparison to the samples described below.
Localised, poorly-developed fabrics were observed but showed no preferred macro-scale
orientation within the sample. Pores are generally isolated and blue-dyed resin has failed
to impregnate some of these (Plate 2).
Fine-grained phenocrysts of pyroxene-type
composition were observed. Total porosity by PIA was S S % for this sample.
3.3.2. SMKS - Pumice
This sample contained clear and generally colourless glass with little evidence of
devitrification or post-depositional alteration (Plate 3). Minor discoloration of the glass
matrix was observed, commonly in areas where vesicles are densely packed with very thin
WG/99/ 15
pore walls.
0 I/ 12/99
Version 1.0
Porosity is dominated by vesicles of 0.02-0.20mm in diameter with
maximum vesicle size around 1.5 mm. Flattening and distortion of the vesicles is greater
in SMK5 than in SMK3 with a curvilinear fabric developed. Infrequent 0.2-0.4 mm size,
colourless phenocrysts of low-relief, simple twinning with first-order birefringence and
inclined extinction were observed. A small number of these phenocrysts are rhombic in
cross-section and possess an oblique twin plane. The fabric adjacent to the phenocrysts
suggests they may have acted as nucleation centres for vesicle formation.
BSEM
identified these phenocrysts as octahedral fluorite (Plate 4) and sanidine crystals (Plate 5).
Localised distortion of pumice fabric is evident in proximity to the phenocrysts with
vesicles adjacent to the grains being enlarged with morphologies that are more complex.
Porosity within the sample was 32 %, significantly lower than observed in the previous
sample.
3.3.3. SMK6 - Muiella pumice
The glass matrix in this sample showed little evidence of devitrification and was
predominantly clear with only a minor localised brown discoloration.
Pore-size is
bimodal. Vesicles of 0.02-0.1 mm size predominate with sub ordinate large vesicles of
0.4-1.2 mm in diameter. Vesicles are highly elongate and exhibit a strong preferred
orientation. A moderate- to well-developed flow-banded fabric is evident throughout the
sample. Total porosity for the sample was 52 % by PIA. Lath-shaped phenocrysts ( ~ %1
by volume) were observed up to 1.5 mm in length. The major axes of these laths are
aligned to the preferred orientation of the fabric. Vesicles adjacent to the phenocrysts are
slightly deformed (Plate 6). Optically these phenocrysts are of similar relief, twinning and
birefringence to those seen in the previous sample and exhibit a 5-10' extinction. SEM
observations identified these crystals as alkali feldspar, most likely sanidine. Angular
fragments were also observed and initially identified as feldspathic xenolith optically.
SEM also identified fragmented, silicified feldspar grains in which with glass has
penetrated along cleavage planes and fractures.
WG/99115
Version 1.0
0 1/12/99
3.3.4. S M K 8 - Pumice lupilli
The glass matrix in this sample showed a slight brown coloration with no evidence of
devitrification. Total porosity by PIA was 66 % with vesicles ranging from 0.01-10 mm
in size with the range 0.3-0.6 mm dominating. Distortion of vesicles was observed with
moderate flattening and elongation within a well developed flow-banded fabric. The
larger vesicles present were observed to be surrounded by regions of high deformation
whereas finer vesicles appear to have been preferentially distorted (Plate 7).
Minor phenocrysts present within this sample were typically 0.2-0.3 mm in diameter and
exhibited low birefringence, rhombic section and compositional zoning (Plate 8).
Textures observed optically showed evidence of alteration andor resorption with euhedral
replacement of the core in some cases. Fragmented grains, up to 10 pm in size, were
observed with parallel and radiating micro-laths present within their structure (Plate 9).
These fragmented grains were identified by SEM as altered plagioclase feldspar, probably
albite originally. Micro-laths observed along original cleavage planes and micro-fractures
within the grains were identified as Mg, Fe-rich phases, possibly clay. Trace quantities of
fine-grained apatite and titaniferrous magnetite were also observed.
3.3.5. SMTI - Airfall pumice
Optically the glass in this sample was generally colourless with fine plagioclase
crystalli tes visible at high magnification. These are well-developed and may represent a
quench texture resulting from rapid cooling of the pumice after ejection.
Maximum
vesicle size observed was about 3-5 mm with the majority of vesicles in the range
0.1-0.3 mm. Total porosity by PIA was 66 %. The phenocryst assemblage in this sample
is more varied than seen previously though still only a minor component volumetrically.
Euhedral augite crystals up to 0.4 mm in length, rutile, quartz aggregates up to 0.7 mm
across, equant opaques, altered feldspar fragments (as seen in the previous sample) and
tabular grains up to 0.25 mm in length were identified optically. SEM identified the
opaques present as predominantly sphene with minor quantities of spine1 (Fe, Ti
dominated) and titaniferrous magnetite. Tabular crystals identified optically as a separate
phenocryst phase are of clinopyroxene composition being rich in Fe, although grains with
enhanced AI concentrations were occasionally observed corresponding more closely to
WG/99/ I5
Version 1.0
0 1/12/99
amphibole composition. Plate 10 shows a shattered, euhedral clinopyroxene (possibly
augite) crystal with a tabular crystal in close contact.
3.3.6. SMT2 - Welded scoriu. scoria
CO~W
Optically this sample is characterised by a dark coloured glass, ranging from sepia
through brown to near opaque (Plate 11) with thick vesicle walls in contrast to the
previous samples. SEM petrography showed extensive quench textures to be present with
well-developed crystallites present within the glassy matrix (Plate 12). EDXA data
obtained from the crystallites showed these to be enriched in Ca and Fe relative to the
glassy matrix. Compositional data for the glassy matrix showed it to have an alkali
composition (Na, K, AI, Si, 0) with only minor Ca and Fe present. Plagioclase and
clinopyroxene, predominantly augite, were observed optically with trace amounts of
apatite and opaques (identified as Fe, Ti spine1 by SEM-EDXA). These minerals were
observed to be present both as single phenocrysts and as micro-phenocrysts within the
glassy matrix.
The porosity of this sample is higher than seen previously, 73 % by PIA, and is poorly
connected with the blue dye failing to penetrate many of the vesicles. The vesicle size is
approximately 0.05-6 mm but, in contrast to the previous samples, is skewed toward the
coarser sizes. Vesicle morphologies observed were typically non-equant but rounded in
contrast to the highly elongate and distorted forms seen in some of the previous samples.
No fabric or preferred orientation could be determined from vesicle orientation or
morphology.
3.4. Physical Properties
Note that generally the quantity of sample available was insufficient for compliance with
British Standard and ASTM tests. The amounts of sample available in comparison to
ASTM requirements are listed in Table 12. To some extent, lack of material has affected
the quality of results obtained.
3.4.I . Sieve unulysis
Frequency distribution and cumulative distributions for each of the samples are shown in
Tables 8 and 9 along with typical results for some commercial examples. Sieve anlysis
WG/99/ I 5
Version 1.0
01/12/99
results are also shown graphically in Figures 6 and 7. It is apparent that SMKS is the
finest sample while SMT2 is the coarsest sample. Inspection of Figure 7 reveals that the
samples can be broadly classified as coarse lightweight aggregate. Fine aggregate could
therefore be obtained from all these samples by mechanical grinding.
3.4.2. Unit weight, water ubsorption, density detemination and voids percentage
All physical property results obtained are summaried in Table 13. Data for commercial
materials are included for comparison. Generally on an ungraded or graded basis, the unit
weights of the East African samples meet the unit weight requirement for coarse
lightweight aggregate for concrete specified by the ASTM (Tables 10 and 11). However,
the unit weight of the scoria sample SMT3 is greater than that specified by ASTM. Also,
the unit weight of the size fraction -1.8-0.6 mm in SMTl is a anomalously high due to the
presence of phenocrysts in this size fraction.
Unit weight is essentially controlled by a combination of three variables: porosity,
particle-size and density of the volcanic glass. For example, the scoria sample SMT2, has
a far lower unit weight compared to the other scoria SMT3. This is primarily because
sample SMT2 is much coarser containing over 50% in the > 14 mm size fraction
compared to less than 2% > 14 mm for SMT3 (Table 8).
SMK6, SMK8 and SMTl are all characterised by low unit weight and high water
absorption which confirms their lightweight, porous nature. In contrast, the scoria sample
SMT3 is characterised by high unit weight and low water absorption. The unit weight and
water absorption of SMKS and SMT2 are intermediate.
Water absorption gives an indication both of total porosity and the degree of the
connectivity between pores. Samples with high water absorption are likely to have higher
porosity andor high connectivity.
The porosity determined by PIA method generally
correlates with water absorption data. For example SMKS and SMTl both have high
water absorption and high porosity.
WG/99/ 15
Version 1.0
0 I/ 12/99
3.4.3. Specific gravity
All the pumice samples have a SG around 2.5, which is similar to the commercial
examples given (Table 13). Unsurprisingly, the scoria samples, SMT2 and SMT3 have
notably higher SG values of 2.81 and 3.03, respectively.
3.4.4. Aggregate impact value (AZV)
AIV results for the samples ranged between 69 and 92% (Table 13). According to the
British Standard the results should be treated with caution if they exceed 30%. AIV is a
procedure designed for testing hard rock aggregate for roadstone and is not a suitable test
for lightweight aggregates. However, this procedure was included to obtain information
on the relative strength of samples. Pumice sample SMKS exhibited the lowest AIV
value (69 %). This sample was also the least vesiculated of those examined, having a
porosity of 32 %.
4. CONCLUSIONS
Of the five pumice samples examined, using a classification based on silica content
(Best, 1982), SMK3, SMKS and SMK6 are of acid composition (>66% Si02) while
SMK8 and SMTl are of intermediate composition (52-66% SiO2).
In comparison,
commercial pumice typically contain 70-71% Si02. The iron contents of the five pumice
samples (4-5% Fe,Ol) are also more elevated than those of commercial pumice
(I-2% Fe,O,).
Of the two scoria samples examined, SMT2 is of basic composition
(45-52% Si02) whereas SMT3 is of ultrabasic composition ( ~ 4 5 %Si02). According to
the discrimination chart of Cox et al. (1979) which utilises silica content and alkali
content as ordinates, SMK3, SMKS and SMK6 can be classed as rhyolites, SMK8 as a
trachyte, SMTl as a phonolite, SMT2 as a phonolitic nephelenite and SMT3 as a
basaniteltephrite.
Vesiculated, poorly-ordered volcanic glass (sideromelane) predominates in all samples.
Minor felsic phases occur as phenocrysts in the pumice samples. In contrast, the two
scoria samples are typified by the presence of mafic phases including clinopyroxene and
olivine. In pumice samples, vesicles are either ovoid, elongated or very flattened in
shape. The maximum vesicle size in sample SMK8 is anomalously large (10 mm) while
WG/99/ 15
Version 1.0
0 1112/99
the porosity of sample SMKS is markedly low (32%). In one of the scoria samples,
(SMT2), vesicles are more equant and rounded than those observed in pumice samples.
In comparison to the ASTM particle-size specification for lightweight aggregate for
concrete, the majority of samples are classed as coarse aggregates. However, SMK 5 is
appreciably finer than the other samples and corresponds more closely to the ASTM size
class for a combined coarse and fine aggregate. Encouragingly, on an ungraded basis, all
samples (except the scoria SMT3) exhibit unit weights within the ASTM specification for
coarse lightweight aggregate for concrete (4380 kg/m’). Pumice samples SMK6, SMK8
and SMTl exhibit low unit weight and high water absorption values indicative of a
lightweight, porous nature. In contrast, the scoria sample SMT3 is notable for its high
unit weight and low water absorption. Of all the samples examined, the unit weight of
sample SMK S most closely matches that of commercial Lipari pumice from Sicily.
5. RECOMMENDATIONS
Further end-use testing is recommended, particularly for the rhyolitic pumices from
Kenya (SMK3, SMKS and SMK 6). Some suggested additional tests are listed in
Table 14. However, any further laboratory testwork is likely to necessitate collection of
further material.
The large quantities of material required for compliance for ASTM unit weight and sieveanalysis methods suggests that physical property tests for lightweight aggregate should
ideally be carried out in the country of origin.
This option is being pursued by
development of field tests luts for unit weight and sieve analysis. The practicalities of
establishing laboratory facilities in the collaborating countries for testing lightweight
aggregate should also be investigated.
6. REFERENCES
ASTM. 1990. Annual book of ASTM Standards Section 4 Construction. 4.02 Concrete
and aggregates.
ASTM. 199.5. Annual book of ASTM Standards Section 4 Construction. 4.01 Cement;
Lime; Gypsum.
WGl99l1.5
0 1/ I 2/99
Version 1.0
Best, M. 1982. Igneous and metamorphic petrology.
BS. 1995. Testing aggregates Part 2: Methods of determination of density. BS 812-
2:1995.
BS. 1995. Testing aggregates Part 112: Methods of determination of aggregate impact
value (AIV). BS 812-112:1990.
BS. 1998. Tests for mechanical and physical properties of aggregates Part 3:
Determination of loose bulk density and voids. BS EN 1097-3:1998.
Cox, K., Bell, J., & Pankhurst, R. 1979. The interpretation of igneous rocks.
Geitgey, R. 1994. Pumice and volcanic cinder. In D. D. Carr (Eds.), Industrial Minerals
and rocks (pp. 803-8 13). Society for Mining, Metallurgy and Exploration, Inc. Litleton,
Colorado.
Harben, P., & Kuzvart, M. 1996. Pumice and scoria. In P. Harben & M. Kuzvart (Eds.),
Industrial Minemls, Global Geology (pp. 3 17-323). Industrial Minerals information Ltd.
London.
Hatch, F., Wells, A., & Wells, M. 1961. Petrology of the igneous rocks
Hyndman, D. 1985. Petrology of igneous and metamorphic rocks.
Inglethorpe, S., Mithchell, C., & Pearce, J. 1997. Characterisation of natural zeolite from
Beli Plast, Bulgaria, and pumice from Sicily, Italy.
British Geological Survey.
WG/97/34C.
Lapidus, D. 1990. Dictionary of Geology.
Loughborough, R. 199 1. Minerals in lightweight insulation. Industrial Minerals October,
21-35.
Melendez, S. A. 1997. Pumice; a rock for all seasons. Mining Voice (Washington, D.C.),
3(4), 11.
Mitchell, C., & Bloodworth, A. 1989. Evaluation of pumice from Costa Rica as
lightweight aggregate. Technical Report British Geological Survey. WG/89/21R.
WG/99/15
Version 1.0
0 1/ 12/99
Robbins, J. 1984. Pumice- the ins and outs of a reinforced market. Zndustriul Mineruls
May, 31-51.
7. ACKNOWLEDGEMENTS
Th authors would like to thank the contributions made to this report by the Analytical
Regional Chemistry Group (XRF and Sample Preparation Facility).
WG/99/ 15
Version 1.0
0 1112/99
TABLE 1. SAMPLE LIST AND DESCRIPTIONS AS SUPPLIED BY THE COLLECTOR.
COLLECTOR
MPG CODE
DESCRIPTION
PUMICE
PUMICE LAPILLI
TANZANIA
WELDED SCORIA,
SCORIA CONE
SMT3
E097
1
I
TANZANIA
1
1
“GOOD’ SCORIA,
FINE, STRATIFIED
N.B. Throughout the report the collector codes are used as opposed to the MPG codes as requested by the
collector.
WG/99/15
Version 1.0
01/12/99
TABLE 2. USES OF PUMICE (AFTER GEITGEY, 1994)
FORM
I
PROCESSING ESSENTIAL PROPERTIES
decorative and structural concrete blocks
cast concrete
H
lightweight structural members
8
;all panels
sga $a
crushing,
sci-eening,
granular
oor decking
s t i i c c ~and plaster mixes
3
low density, good crushing
strength, thermal insulator,
acoustical insulator
pozzolan in cement
civil engineering
lightweight fill
grill cleaners
scouring sticks for porcelain, tiles and swimming pools
blocks
sawing irregular
lumps as mined
broken vesicle (bubble) walls
form sharp edged particles;
wear continues to generate
fresh cutting edges
coarse granular
crushing,
screening
as above
granular
drying, milling,
screening, air
flotation, blending
broken vesicle (bubble) walls
form sharp edged particles;
wear continues to generate
fresh cutting edges
granular
crushing,
screening
high porosity, large surface
area, low chemical reactivity
coarse granulai
crushing,
screening
no information
gi-anular
crushing,
screening
low density, thermal insulator
acoustical insulator, fire
resistance, moisture resistancc
blocks, boulder
as mined
sawn slabs
low density, easily shaped,
low maintenance
crushing, drying,
milling, screening
blending
particle shape, cost
buffing wheel cleaners
cosmetic skin removal
stonewashing
~
r/)
Y
2
3
4
~
hand soaps
scouring compounds
rubber erasers
polishing compounds for glass, metal, plastics
dental cleaners
wood finishing
nonskid paints
cleaning printed circuit boards
tumble polishing
leather finishing
matches and striking surfaces
potting soils
2
hydroponic media
pct litter
floorsweep
U
Y;
turf aeration
4
acld washing
rr
catalyst cairiers
,
5
loose till insulation
w
roofing granulcs
textured coatings
W
ground cover
5
0
E
c
landscaping
'
rubber, paints and plastics
k
hot asphalt mixes
brake linings
Y
$3
58
E 2
I
mold release compounds
granular
I
crushing, drying,
particle shape, expandability
animal, vegetable and mineral oils
firing, air flotatior
1
3
0
m
w
9
c
.-0
ii
n
8
iz
4
4
u
p:
EE
0
U
w
E
5
od
-
0
rn
nz
m
9
0
l
o
\
CI
3
3
\4
m
0
n
z
-
w
r
Q
0
0
r?
-?
c?
3
c
rn
W
x
o
\
-
n
z
m
CI
s
09
3
3
~
ln
m
x
d:
In
3
3
m
x
ln
k4
0
0
p!
w
Lc!
0
r-
m
8r-
WG/99/ I5
Version 1.0
0 1112/99
% (AFTER
d
0
'"
r-
00
m
09
0
0
0
8
2
8
rc\1
09
hl
3
W
x
UI
m
9
0
hl
c?
2
TABLE 6. SUMMARY OF XRD RESULTS
Sample
Mineral phases present in approximate order of concentration
SMK3
Glass, feldspar (undifferentiated), quartz,?cristobalite
SMKS
Glass, quartz, a1kali feldspar, plagiclase
SMK6
1 Glass, feldspar (undifferentiated), quartz
SMKS
Glass, feldspar (undifferentiated), quartz
SMTl
Glass, alkali feldspar, plagioclase
SMT2
Glass, clinopyroxene
I Glass, olivine, clinopyroxene
SMT3
TABLE 7. DESCRIPTIVE DATA FROM OPTICAL PETROGRAPHY
~~
~
Sample
~
Porosity
Average pore size
Maximum pore size
(%)
(mm)
(mm)
P
Pore shape
Pore connectivity
~
Sub-spherical to
SMK3
90
SM K5
50 - 60
SMK6
80 - 90
I
I
0.02 - 1.5
0.02 - 0.1
I
1
Flattened
0.4X1.2
Elongated
70
0.05 - 6.0
NIl=no data
N B A thin section was not prepared for SMT3
I
I
Poor
Poor
Y-+-Flattened
3.0 - 5.0
SMTl
Poor
ovoid
I .5
0 . l X 0.3 - 0.6
SMKS
SMT2
3
0.1 - 0.5
ND
Rounded
Poor
.
3
0
-
3
3
CO
N
d
a
N
- h
E
E
v
x+
d
Q)
m
+
W
+
v)
?
\o
c?
N
z
9
v,
2+ a
+
T
!
3
v)
4
c?
4
'4
c!
r;'
n
z
'c!
m
c?
c!
c!
c?
*
CO
N
c?
*
3
3
0
3
3
m
3
3
m
3
3
3
3
2
z
c?
d
09
c?
i
b,
?
a
'c!
r-
c?
m
N
-
m
N
F?
v!
z
9
3
m
c!
\?
3
N
i
P
m
3
8
- -?
10
9
3
0
3
9
3
x
9
i
P
9
2
x
?
m
3
3
3
3
3
c?
d
09
0
c!
0
c!
c?
c?
c?
0
0
3
hi
-?
0
_
.
9
m
m
3
0
?
m
3
'c!
?
2
3
N
U
+
3
+
v,
E;
W
c?
CI
+
s
L
x
0
0
i
hi
9
m
i
3
m
s--.
In
W
3
n
0
d
m
3
3
0
x
3
r-:
hl
F:
ErL
00
m
Pi
hl
.
4
0
Q,
N
-
.e
m
P)
U
.
I
U
L
8
0
d
P
d
m
vi
d
d
d
0
r-
vi
3
b
e
t?
1%
if
a
m
a
\o
a
2
d
0
d
a
m
m
00
r-
Q
m
d
CO
d
vr
9
3
c
.-0
L
m
>”
15
n
z
z
Y
d
m
00
15
3
vj
E:
0
Q
N
hl
CI
vi
r-
00
m
m
L+
m
m
15
Y
E
CA
1N
0
\o
N
In
hl
m
In
rm
In
+
0
w
Q)
N
u
p!
4
L+
CA
m
4
hl
.L
m
s
vi
r-
00
P
d
9
3
0
0
2s 20
a
z
8
lT
2
-
G
m
ja
v
l
109
h l v l
1
d
'c!
c\I
3
E
v)
0
'c!
hl
m
b.
hl
m
Ld
E
v)
m
+
I--
WG/99/15
Version 1.0
0 1112/99
TABLE 14. ASTM TEST METHODS FOR LIGHTWEIGHT AGGREGATE
BGS
ASTM
PROCEDURE
NUMBER
C29
CAPABILITY
Test method for Unit Weights and Voids In Aggregate
.......................................
c39
Test Method for Compressive Strength of Cylindrical Concrete
N
Specimens
.....................................
C40
.............................................................................................................................................................................................................
Test Method for Organic Impurities In Fine Aggregates for concrete
.......................................
C114
Test Method for Chemical Analysis of Hydraulic Cement
Y (L.O.1)
.............................................
.......................................
C136
Y
.............................................
Method for Sieve Analysis of Fine and Coarse Aggregates
Y
Test Method for Clay Lumps and Friable Particles in Aggregates
Y
.......................................
C142
......................................
C151
.............................................................................................................................................................................................................
Test Method for Autoclave Expansion of Portland Cement
Cl77
N
.............................................
......................................
Test method for Steady-State Heat-Flux Measurement and Thermal
N
transmission Properties by Means of a Guarded Hot Plate Apparatus
......................................
C289
.............................................................................................................................................................................................................
Test Method for Potential Reactivity of Aggregates (chemical Method)
C295
Y
..............................................
......................................
Practice for Petrographic Examination of Aggregates for Concrete
Y
Test Method for Compressive Strength of Lightweight Insulating
N
......................................
c495
Concrete
......................................
C496
...................................................................................................................................................................................................
Test Method for Splitting Tensile Strength of Cylindrical Concrete
N
Specimens
.............................................
......................................
C513
Test Method for Obtaining and Testing Specimens of Lightweigh1
N
Insulating Concrete
.............................................
.....................................
C64 1
Test Method for Staining Materials in Lightweight Concrete Aggregates
C666
.....................................
C702
Y
.............................................
.....................................
Test Method for Resistance of Concrete to Rapid Freezing and Thawing
N
.............................................................................................................................................................................................................
Practice for Reducing Field Samples of Aggregate to Testing Size
Y
Download